ABSTRACT Scale is a fundamental obstacle in linking neural activity to behavior. While perception and cognition arise from interactions between diverse brain areas separated by long distances, neural codes and computations are implemented at the scale of individual neurons. An integrative understanding of brain dynamics thus requires cellular-resolution measurements across sensory, motor, and executive areas spanning more than a centimeter. While such comprehensive sampling has recently been achieved in head-fixed animals, comparable data during complex behaviors in freely moving animals is not possible with current technology. Head-mounted ?miniscopes? provide powerful tools for long-term monitoring of neural population activity in behaving mice. While new surgical methods now offer optical access to most of neocortex, current miniscope optics limit their field of view (FOV) to less than 1 mm2, precluding imaging across multiple brain regions. Large- scale imaging also requires mouse models with widely expressed Ca2+ sensors, severely constraining size and weight. Fundamental physical principles dictate that conventional optics cannot meet the joint requirements of large FOV, high resolution, and efficient light collection in a compact and lightweight device that can be carried by small animal models. Here we propose to develop a next-generation miniscope for cellular-scale imaging across the majority of cortex in freely moving mice, based on novel computational imaging framework that jointly design optics and algorithms. Our Computational Miniature Mesoscope (CM2) uses parallel sampling with single lightweight microlens array to dramatically simplify the optical path, increasing FOV by ~2 orders of magnitude while maximizing light-throughput, resolution and signal-to-noise ratio. The CM2 system will provide resolution of approximately 10 m across distances of centimeter or more, enabling brain-wide recording of multiscale neural dynamics inaccessible with existing technology. Our goals are to: 1) Develop novel optical designs for ?wearable? cellular-resolution Ca2+ imaging throughout neocortex. 2) Validate and optimize the CM2 for cortex-wide functional imaging in behaving mice. To this end, our interdisciplinary team brings together highly complementary expertise in computational imaging, neurophotonics innovation, and neural recordings in behaving rodents. Overall, this work will establish powerful enabling technology that greatly expands the scale of activity measurements possible in behaving animals, providing access to a wide range of questions about distributed cortical function.